Current methods of summarizing data from epidemiologic studies of periodontal disease (such as the PI,k PDI,k and CPITN) reduce detailed clinical examination data to one score that summarizes the severity of disease. The purpose of this project is to investigate the possibility of a different summary statistic which will better characterize the clinical data. The data used for analysis had been collected previously during a longitudinal study of periodontal disease. There were 369 males, age 17-32 years, examined twice, three years apart. The data consisted of 56 attachment loss measurements taken at the mid-buccal and mesio-buccal sites of all teeth excluding the third molars. Criteria were established defining a "diseased" site and a "diseased" subject. Computer programs were written to compute different disease parameters for this population. From these analyses, a bi-variate statistic was derived called the Extent and Severity Index (or ESI), written (E,S). The ESI summarizes the average extent of sites diseased within the mouth (E) and the average severity of disease in millimeters (S). Without sacrificing accuracy, it was shown that a half-mouth examination design can be used (mid-buccal and mesio-buccal sites on the second molar through the central incisor, random assignment of two quadrants) which reduces examination time. E and S provide independently useful information, and can be tested with multivariate statistical procedures (such as Hotelling's T2) to determine significance over time or between groups. The technique and rationale for computing ESI was presented at a conference on periodontal disease, and will be published in the Journal of Clinical Periodontology in 1985-1986.